Testing and Modelling of a Severely Tapered Composite Specimen

Stephen Hallett, James Lander, Mike Jones, Luiz Kawashita

and Michael Wisnom

www.bris.ac.uk/composites

2Introduction• A severely tapered specimen has been

designed and manufactured at the University of Bristol

• Purpose is to be representative of aerospace component features

• Made up from a large number of pre-preg plies dropped off to create thickness change

• Tested in a dovetail type fixture• Purpose if to inform and validate high

fidelity modelling of situations where failure is governed by delamination from ply drops

3Summary• Overall specimen configuration:-

• Specimen manufacture• Static Test Results

– Failure load/location– High speed photography– DIC– CT Imagery

• Modelling

Overall Length = 310 mmGripped length = ~140 mmWidth =20mm

~11.2mm~39.2mm

4Manufacture

Key Requirement: Accurate, reproducible positioning of ply drops

5Laying Up

E-glass/914 Contact Pads laid down first

IM7/8552 laid directly on top

Co-cured

6Continuous Visual Accuracy AssessmentFull thickness

Increasing thickness

7Edging

Edged with Neoprene doped adhesive cork strips

Contour edging detail

8Autoclave cure

9Finished Part

• Specimens were cut from cured plate

• Each was 20mm wide

• Surface scan shows well placed plies with good consolidation

10Test Fixture

• For static (tensile) and fatigue (tension-tension) testing

• Quarter fixture FE analysis performed

• Abaqus generated von Mises stress plot• Steel fixture

• Titanium inserts

11SRP Static Tensile Testing Setup

SRP Test Specimen

High Speed Video Camera

Load Cell

SRP Test Fixture

12Test Curves - Load vs. Displacement

c.v. = 8.8%

13Test Curves - Close-up of failure region

specimen p1_2(the only specimen taken to ultimate

catastrophic failure)

=first failure

14Ply Drop Map

3

2

45

1

15Failure LocationWhite markers indicate PD 1, PD 2 and PD 5

Initiation occurs above PD 5

16

ply drop #4

High Speed Camera Video – specimen p1_8

Frame rate: 360,000 fps

45 345 2345

12345

45 345 2345

12345

17

• Random black on white speckle pattern applied to specimen using spray paints

• 2D system used as no significant out-of-plane deformation expected

• Images captured at 2 sec intervals throughout test (i.e. approx 450-500 images per test)

• Images processed post-test using Dantec Istra 4D software

Digital Image Correlation

High contact strains

Failure location

96.3% Failure load

99.8% Failure load

18CT Scans

PD 4

Arrangement is not to scale

A

C

B

PD 1

B

A

C• CT scan was taken of representative sample

• Shows voids at some critical ply drop locations

19High-Fidelity FE Modelling• Methodology based on the use of custom (user-defined) cohesive

elements for modelling delamination in the presence of high through-thickness compressive stresses

• Cohesive interfaces defined between every ply, as well as around resin pockets and glass pads

• Required the development of automated meshing software to tackle the complexity of such models

• Meshing tools used to provide realistic specimen and ply drop geometries

19

20

• Basic procedure builds a 3D model from a 2D drawing of specimen profile and layup

High-Fidelity FE Modelling 20

21

• In house meshing tool builds a 3D model from a 2D drawing of specimen profile and layup

High-Fidelity FE Modelling

resin elements

cohesive elements

-45°

0°

+45°

• Plies, resin pockets and cohesive elements are all generated automatically• Creates realistic ply terminations (can be fine-tuned to match real laminate)

21

22Test/Model Correlation: Failure Location 22• FE results were used to predict the site of delamination initiation – useful for

setting up the high-speed camera• 5 out of 6 specimens delaminated within 1-2 millimetres from predicted

locationP

delamination (shown in green)

2323Test/Model Correlation: Load-Displacement• Force-displacement curves for (2D) slice and 3D models

2424Effect of Ply Drop Geometry

idealisedply drops realistic

ply drops

‘idealised’ ply drop

‘realistic’ ply drop

25Effect of Friction

m = 0.3

m = 0.0m = 0.1m = 0.2

26Effect of Voids

1 2

3

No voids

1 2

27Summary

• Testing has been undertaken on a severely tapered specimen with multiple ply drops

• Failure occurs in the region ahead of the contact area where delamination is not suppressed by compression

• Various experimental techniques have been used to identify failure initiation location

• Finite element analysis has been able to predict very similar behaviour

• Model results are very susceptible to small details; resin pocket geometry, friction, voids which need to be captured accurately from experimental work

28

The authors would like to acknowledge Rolls-Royce plc for funding this work

Acknowledgements

Slide Number 1IntroductionSummaryManufactureLaying UpContinuous Visual Accuracy AssessmentEdgingAutoclave cureFinished Part Test FixtureSRP Static Tensile Testing SetupTest Curves - Load vs. DisplacementTest Curves - Close-up of failure regionPly Drop MapFailure LocationSlide Number 16Digital Image CorrelationCT ScansHigh-Fidelity FE ModellingHigh-Fidelity FE ModellingHigh-Fidelity FE ModellingTest/Model Correlation: Failure LocationSlide Number 23Effect of Ply Drop GeometryEffect of FrictionEffect of VoidsSummaryThe authors would like to acknowledge �Rolls-Royce plc for funding this work